Who Is “Mass Online Surveillance” for?

• In Event: Free Speech, Censorship, & Self-Censorship

Abstract

Post dot-com bubble, mass adoption of digital technologies aided by cheaper devices and computational advances, has spurred the growth of a business model that has been dubbed “surveillance capitalism” (Zuboff 2015). Selling consumer data to marketers and advertisers is not new; this model makes analysis of individualized consumer behavior possible by keeping data collection across devices and matching online to offline behavior in a way previously impossible. Alphabet (Google’s parent company), Microsoft, Facebook, and Amazon are incentivized by the logic of surveillance capitalism to create platforms, services, and devices that collect, harvest and connect individual people’s personal data on a mass scale. These companies aim to “nudge” consumers to make purchases (IBM 2017; Varian 2014; Zuboff 2015). Why have these tactics remained unregulated, particularly in America, where these companies are headquartered? This paper builds on Schneier’s argument that governments have little incentive to prevent private companies from engaging in online surveillance and data collection in part because government agencies rely on by these companies to reduce the cost of their own surveillance (2015). In exchange for allowing companies to collect user data with little oversight, government agencies are granted almost unlimited access to the data collected by private firms with little transparency.

While these trends are raising concerns in fields such as computer science (e.g., Buolamwini and Gebru 2018) and business (e.g., Zuboff 2015), less work in the field of political science considers how mass surveillance impacts citizens’ behavior, and what this means for democracy. There is some research that shows that awareness of government monitoring, not fear of prosecution, depresses online searches for sensitive topics on Wikipedia (Penney 2016). What is more concerning is that these concerns asymmetrically impact the population. Disempowered social group members—including women, low-income citizens, and younger people—are more likely to engage in online self-censorship.

In this paper, I present results from an original survey experiment (n = 1,493) that offers insight into Americans’ understanding of privacy and surveillance, clarifies the conditions under which Americans are willing to tolerate surveillance, especially the role that group-based processes (such as affective polarization) play in increasing citizens’ acceptance of surveillance. Based on a novel, multidimensional measure of Americans’ privacy concerns, I use a 2X3 vignette experiment to show that 1) tolerance for surveillance is higher when people are primed to think of a disliked group member compared to a liked group member, or no member (control) and that 2) even when there is mass data capture and analysis, if a policy’s effect is framed as producing targeted results, willingness increases. I follow up the experiment with an affect counterargument (keeping policy constant within-respondents) and find 3) statistically significant order effects. Tolerance for surveillance of disliked groups can be lowered if liked groups are primed first, but not if they are mentioned in the counterargument. The control group is randomized into positive/negative affect conterarguments and respondents behave expected. Tolerance decreases for liked group negative externalities but increases for disliked groups, in reverse of those who were treated with explicitly opposite affects.

The implications of my work for the field of political behavior are threefold. First, it engages with a salient and fluid policy choice (mass online surveillance) that has been understudied in the context of advanced democracies. Second, the experiment expands the application of sociotropic/egotropic dynamics beyond economic voting. Third, the paper builds on affective polarization work to explain tensions and changes in public opinion of policy.

Author

• Isadora Borges Monroy, McGill University